Life expectancy gap between the Francophone majority and Anglophone minority of a Canadian population

Abstract

Language is an important determinant of health, but analyses of linguistic inequalities in mortality are scant, especially for Canadian linguistic groups with European roots. We evaluated the life expectancy gap between the Francophone majority and Anglophone minority of Québec, Canada, both over time and across major provincial areas. Arriaga’s method was used to estimate the age and cause of death groups contributing to changes in the life expectancy gap at birth between 1989–1993 and 2002–2006, and to evaluate patterns across major provincial areas (metropolitan Montréal, other metropolitan centres, and small cities/rural areas). Life expectancy at birth was greater for Anglophones, but the gap decreased over time by 1.3 years (52% decline) in men and 0.9 years (47% decline) in women, due to relatively sharper reductions in Francophone mortality from several causes, except lung cancer which countered reductions in women. The life expectancy gap in 2002–2006 was widest in other metropolitan centres (men 5.1 years, women 3.2 years), narrowest in small cities/rural areas (men 0.8 years, women 0.7 years), and tobacco-related causes were the main contributors. Only young Anglophones <40 years in small cities/rural areas had mortality higher than Francophones, resulting in a narrower gap in these areas. Differentials in life expectancy favouring Anglophones decreased over time, but varied across areas of Québec. Tobacco-related causes accounted for the majority of the current life expectancy gap.

Appendices

Appendix 1. Statistical description of Arriaga’s method for decomposing a life expectancy gap

Changes in life expectancy over time (or differences between populations at the same time) are a function of changes (or differences) in age-specific mortality. For each age group, Arriaga’s method estimates how many years of life expectancy at birth are added (or removed) due to change (or difference) in age-specific mortality rates. For example, the effect ( _n ∆ _x , in years) of mortality change from 1990 to 2001 between ages x and x + n on the change in life expectancy at birth is calculated as

$$ {}_{n}\Updelta_{x} = \left[ {\frac{{l_{x}^{1990} }}{{l_{0}^{1990} }} \times \left( {\frac{{{}_{n}L_{x}^{2001} }}{{l_{x}^{2001} }} - \frac{{{}_{n}L_{x}^{1990} }}{{l_{x}^{1990} }}} \right)} \right] + \left[ {\frac{{T_{x + n}^{2001} }}{{l_{x + n}^{2001} }} \times \frac{{\frac{{l_{x}^{1990} l_{x + n}^{2001} }}{{l_{x}^{2001} }} - l_{x + n}^{1990} }}{{l_{0}^{1990} }}} \right] $$

(1)

where l _x is the number surviving to age x out of a synthetic cohort (l ₀ is the cohort size, typically 100,000 in a period life table), _n L _x is the number of person-years lived between ages x and x + n, and T _x+n is the total number of person-years lived above age x. The first term on the right of Eq. 1 is the direct effect of mortality change between ages x and x + n, or the change in the number of years lived between ages x and x + n. The direct effect is the product of the fraction of the cohort surviving to age x in 1990 (l _x /l ₀) multiplied by the change in the average number of person years lived from ages x to x + n ( _n L _x /l _x ) between 1990 and 2001. There is also an indirect effect of mortality change within an age interval because the direct effect produces additional survivors at the end of the interval, and an additional interaction effect due to the additional survivors who are exposed to new mortality conditions in the later period [17]. The second term on the right of Eq. 1 captures the indirect and interaction effects, and is the product of the life expectancy remaining at age x + n after the mortality change (T _x+n /l _x+n in 2001) and the fraction of additional survivors in the birth cohort [18]. For the last age group (90 + years), there is only a direct effect of mortality change. The decomposition between language groups in a given year is obtained by replacing year in Eq. 1 (1990, 2001) with language (Francophone, Anglophone).

In addition, the overall contribution of each age group may be partitioned by cause of death under the assumption that the contribution of each cause to the life expectancy change for an age group is proportional to the contribution of each cause to the change in the overall age-specific mortality rate [19]. The literature suggests this is a reasonable assumption, but estimates of the contributions for common causes of death heavily concentrated at older ages may be conservative [57, 58]. The contribution to mortality change of cause i within ages x and x + n from 1990 to 2001, ${}_{n}\Updelta_{x}^{i}$, is estimated as

$$ {}_{n}\Updelta_{x}^{i} = {}_{n}\Updelta_{x} \times \frac{{\left( {{}_{n}p_{x}^{i,2001} \times {}_{n}r_{x}^{2001} } \right) - \left( {{}_{n}p_{x}^{i,1990} \times {}_{n}r_{x}^{1990} } \right)}}{{{}_{n}r_{x}^{2001} - {}_{n}r_{x}^{1990} }}, $$

(2)

where ${}_{n}p_{x}^{i}$ is the proportion of deaths between ages x and x + n due to cause i, and _n r _x is the all-cause mortality rate between ages x and x + n. Thus the contribution of a change in a specific cause of death is a function of both the absolute change in age-specific mortality ( _n r _x ) and the change in the distribution of the cause over time relative to other causes (${}_{n}p_{x}^{i}$). From Eq. 2, it is inferred that causes of death less frequent in 2001 than in 1990 will make positive contributions to the overall positive change in life expectancy at birth, while causes of death more frequent in 2001 will make negative contributions.

Appendix 2

See Table 5.

Table 5 International classification of disease (ICD) codes for causes of death

Appendix 3

See Table 6.

Table 6 Cause-specific age-adjusted mortality rates according to sex and language spoken at home, Québec, 1989–1993 and 2002–2006